Revision Total Knee Replacement- Second stage using PFC – MBT with metaphyseal sleeve and stem (Depuy)

Ensure all imaging available prior to starting
The imaging must stay up on the screens during surgery. Multiple software packages are available to assist templating implants.
On these pre-operative radiographs, the knee cement spacer moulds are apparent. These were made ex-vivo at the time of the first stage revision and contain both cement and targeted antibiotics. A few additional beads have been inserted down the tibia. There has been no significant bone loss from the time of surgery between stages.

Place leg supports to allow full range of knee movement.The positioners should be placed such that the knee can bend to at least 90 degrees but remember that flexion is decreased in revision operations.
The lateral positioner should be placed over the thigh tourniquet allowing the leg/tibia to remain vertical without ‘flopping’ into varus or valgus.

There are many ways to isolate and prepare the leg prior to surgery.
For me the key factors are firstly that there should not be any skin showing, it should all be covered by a combination of impervious stockinets and iodine impregnated drapes. Secondly, ensure there is room proximally to allow for skin incisions.
The tourniquet is inflated following limb elevation with the knee flexed at 90 degrees to allow some quadriceps movement.

Access starts with incising through the existing scar.If there is a single midline scar then access should be through this scar.
If there are multiple scars the general rule is to use the lateral scar. Vastus medialis extends more distally than vastus lateralis. Therefore the lateral blood supply is the most tenuous. A more medial the incision results in a larger lateral skin flap which is at greater risk of wound complications.
However, the incision used must allow full access to the knee. It is generally stated that the skin bridge between scars should be 7cm

Develop the surgical plane superficial to the deep fascia and extend this both medially and laterally.The incision needs to go down to the deep fascia. This is immediately superficial to the extensor mechanism.
If the surgical plane is not obvious extend the proximal incision to virgin tissue and expose the muscle fibres of vastus lateralis. The plane is immediately superficial to this and can be extended distally.
Do not perform ANY subdermal dissection. The blood supply to the skin is from a subdermal plexus supplied from the descending branch of the lateral circumflex, the genicular (popliteal) and anterior tibial recurrent arteries. Skin is most at risk over patella, tendon and tibial tubercle.
An excellent review of wound problems following total knee replacement can be found in;
Wound Healing Problems in Total Knee Arthroplasty. Garbedian et al Orthopedics, 2011, Vol.34(9), pp.516-518.

Care must be taken not to drift too deep or superficial when developing the deep planes.
When elevating the tissues ensure that the forceps do not clasp skin edges. This increases the risk of wound breakdown and complications.

Aspirate the joint prior to the arthrotomyTo ensure minimal risk of contamination the joint can be aspirated before being opened. Use a white needle and syringe.

The arthrotomy should follow previous incisions to avoid further devascularisation.This usually will be in the midline proximal to the patella (in the ligamentous area of the quads ligament), then medial to the patella (leaving a small cuff for closure) and then most distal, medial to the patella ligament.

Complete the arthrotomy exposure with the leg in extension
Placing the leg in extension decreases the tension in the tissues.

Scar tissue will have formed deep to the knee capsule which needs to be removed once the arthrotomy has been completed.This can be challenging but I start in the midline of the femur, immediately proximal to the femoral component. I strip tissue off the bone both medially and laterally into the medial and lateral gutters.
Laterally identify the joint capsule (A), dissection continues deep to this. Moving laterally scar tissue is removed.

Continue the lateral intraarticular debridement, removing scar tissue.
Eventually, the lateral dissection will ‘meet up’ with the dissection performed over the femur allowing a considerable amount of scar tissue to be removed.
This is a skill that needs to be learned by doing and cannot be taught in a few slides.

Even when an articulating spacer is used at the first stage, a second stage knee replacement is often very stiff.
Forcing the knee into flexion before releasing fully can often lead to significant complications including skin tears, patella tendon pull-off or fracture. It should be avoided.
Therefore patiently resect scar tissue, ensuring full exposure.
Care should be taken posteriorly. To dissect the posterior aspect of the knee elevate the femur with Hohman’s. Take care not to leaver on the tibia too much as this can be deformed if soft. Commercial retractors are available for this. Whilst dissecting posteriorly always pay attention to the type of tissue. A layer of fat surrounds the popliteal arteries. When fat is encountered do not go any deeper.
I have dissected this area in cadaveric workshops. This fat layer is approximately 6mm deep. Take care.

Confirm full exposure before removing implantsAt the end of the soft tissue clearance, the implant should be fully exposed without undue tension on the soft tissues.

Remove the articulating spacer with an osteotome, piecemeal if required.The spacer can be removed with a mallet and osteotome however be cautious not to be overzealous. Only the implant should be removed. The temporary cement spacer is constructed ex-vivo in plastic moulds and antibiotics are added to the cement. It is implanted once cured with a ‘press-fit.’

Once the implant has been fractured with the osteotome, remove the cement.

Take care not to remove or damage underlying bone during spacer removal. There may well be little bone stock.
As can be seen in this image, the bone stock remains acceptable however there is scar / fibrinous tissue over the femur.
This can be gently removed with a curette or Charnley spoon. The bony surface needs to be visualised to allow accurate femoral cuts to proceed.
The tibial tray is apparent in this image too and can simply be removed with an osteotome.

It is apparent from the x-ray that there are intramedullary beads down the tibial canal. These can simply be removed with forceps / Kocher’s. The beads are a means of getting extra antibiotics into the patients. The surface area to volume ratio of beads allows rapid ellution.

Send off laboratory samplesPre-preparing the lab forms is useful as the theatre team can send off samples before the end of the case.
In my institution, ALL revisions have samples taken. 5 samples taken with clean blades and forceps.
The samples are sent off for extended and enriched broth culture. They are prepared in a Class 2 laminar flow hood to protect the samples from contamination.

MBT Reamers and drills are required for initial preparation of the the tibia and femur.
The next two pieces of kit are the 9mm drill (A) and the Hand Reamer (B).
Note that the hand reamer requires 2 adapters. The hand reamer has a number of laser etched depth lines (C). These go from 1-4 and are relevant for the depth of reaming regarding different types of stem.
From the Op Tech;
Press-Fit stems.
Stem Length Reamer Line
75mm 2
115mm 3
150mm 4
Cemented Stems
30mm 1
60mm 2
90mm 2.5
120mm 3.5
150mm 4

Use the initial 9mm drill to open up the medullary canal of the tibiaBefore plunging in the drill use the AP and lateral radiographs to estimate the centre of the medullary canal.
The drill should be advanced no more than 4cm.

Expand and ream the tibial medullary canal initially using the hand reamer, to the required depth.The level at which to ream is most often determined from pre-operative radiographs.
Rhidian Morgan Jones from Cardiff, a renowned revision surgeon, has popularised the theory that stability of the revision prosthesis is conferred by having a good implant/bone interface at the tibial plateau, the tibial metaphysis and the tibial shaft. In an ideal world, all three areas would confer stability but often, with bone loss, this is not possible. Therefore ‘two out of three ain’t bad.’
The tibial stem has to help bypass any poor proximal bone and allow transfer of load from a compromised plateau/metaphysis to the cortical/diaphyseal bone.
The type of stem (cemented versus uncemented) and length remains somewhat controversial. The stem has to be long enough to ensure good bone/implant interface over a significant distance without being too long which increases the risk of stress shielding or implant fatigue fracture.
The summary article by Kang et al (below) is a good starting point when reviewing the use of stems.
Stem Fixation in Revision Total Knee Arthroplasty: Indications, Stem Dimensions, and Fixation Methods. Kang et al Knee Surg Relat Res 2018;30(3):187-192.

Sequentially ream the tibial medulla carefully with the power reamer, after initial hand reaming.In this instance fluted, uncemented stems were to be used. The length of the stem required was assessed using the pre-operative radiographs. A 75mm stem would be adequate as the proximal tibial bone was not insufficient and additional metaphyseal support was planned using metaphyseal sleeves. Therefore from the chart (documented in slide 19) reaming should proceed to the 2nd reamer line.
As a word of caution, the reamer line depth is measured from the planned tibial joint line. Therefore if tibial build-ups are required the reamer line should only go down to the top of the tibial build-up, not to the bone. If you measure from the tibial plateau (and it is deficient) there will be distal over-reaming.
Once the reaming depth is achieved, sequentially increase the reamer diameter until firm endosteal bone is encountered. Bone reamings should be visible in the reamer flutes. If this does not happen then the reamer is under-sized. Reaming increments are 2mm.
The maximum diameter stem that will fit through the metaphyseal sleeve is 14mm Therefore if considering a larger stem note that if revision is required the metaphyseal sleeve will need to be removed before the stem. This is not always straightforward.

Prepare the MBT revision tapered reamer
The tapered reamer is constructed using the MBT tapered reamer plus the stem trial (inserted inferiorly (A) and turned clockwise to secure). The diameter of the stem trial is the diameter of the last used reamer (slide 22)
Note the ‘notch’ on the tapered reamer (B). This is important for assessing tibial resection levels (see later).
This reamer prepares the tibia for the MBT broaches.

The MBT revision tapered reamer is finally advanced until the correct notch is at the level of the tibial resection.Where limited tibial resection is planned this usually corresponds to the tibial plateau, however when the tibial plateau is irregular of there has been significant proximal bone loss the notch may need to be advanced distally into good bone (the reasons for this will be explained further in the forthcoming slides).

Decide whether Tibial Sleeves are going to be used.
The tibial sleeves are used when there are metaphyseal defects.
Engh developed a classification system based on which part of the metaphysis was deficient. 1 – contained defect, 2A – One plateau (medial or lateral) involved, 2B – both involved and 3 – significant loss of proximal bone.
The Op Tech states that sleeves can be used for Type 2 and 3 defects. I would advocate caution when contemplating a tibial reconstruction in which significant proximal tibial bone has been lost. A proximal tibial replacement may be the better option. The decision between and large sleeve and proximal tibial replacement is beyond a few lines on this website and cases are usually discussed between fellowship-trained consultants at formal MDTs.
As you can however see in this slide, the metaphyseal sleeves sequentially increase in size.
The notch on the tapered reamer (slide 23) corresponds to the top of the sleeve broach (A) which corresponds to the level of tibial bone resection. If the tapered reamer is not advanced sufficiently to the laser notch there is a risk of splitting the tibia when the broaches are introduced.

Impact the first sleeve into the metaphysis starting with the smallest broach. The sleeves are asymmetrical, therefore ensure that ‘ANT’ is positioned anteriorly.For small defects align the broach anatomically with mid-anterior part of the broach over the middle third / lateral two-thirds junction of the tibial tubercle. Other reference points are available (tibial spine or centre of the ankle) and all have their shortcomings as reported on by Cobb et al.
For larger defects, it may be that the sleeve is placed in such a way to fill the void. There are 20 degrees of ‘rotational play’ allowed between the sleeve orientation and the tibial tray orientation.
The anatomical tibial axis reliable rotational orientation in knee replacement. Cobb J et al JBJS (Br) 2008:90B

Resect the proximal tibia with the introducer handle removed, the top of the broach can be used as a resection guide.

Remove the tibial broach to complete the tibial resection.

The trial base plate can be used to confirm that the tibial cut is correct, size, rotation and orientation of the cut are all be assessed.

Assemble the Tibial Trial and then construct the tibial impactorThe constituent parts of the trial are assembled. This includes the tibial tray, the tibial stem and the metaphyseal sleeve.

As the tibial defect in this case was minor, the sleeve was orientated anatomically. The centre of the tibial tray corresponds to the centre of the sleeve.
There are other methods to accurately align the tray to the sleeve. These have been described in my earlier case where I used the MBT TKA with sleeves in a tibial plateau fracture (Slide 32.)
Sleeved Total Knee Replacement for Tibial Plateau Fracture (MBT DePuy)
The trial sleeve is inserted into the space created by the broach and gently impacted. The tray and stem are then inserted through the sleeve, allowing independent orientation of the tray on the sleeve (up to a maximum of 20 degrees).
The tibial impactor (A) is used to impact the tibial trial.

Confirm orientation during final impaction of the tibial trialIf the tibial baseplate and sleeve have been correctly orientated the tibial trial assembly will seat down with correct alignment. This needs to be confirmed visually.

Whilst this is not documented in the op tech, it is very useful before embarking on the femoral preparation just to get a simple sense check of the proposed implant on the distal femur.
Does the planned femoral component fit? Is the rotation correct? Can you see any major bone deficits and will they need augments?

Prepare the femur, Initially using the 9mm drill.Repeat the initial reaming for the femur using the 9mm drill. Again ensure that the drill follows the medullary canal. Avoid extension or flexion.

Sequentially expand the medually reaming of the femur.Again ream only to the laser etching that corresponds to the predetermined femoral stem length. The stem length (unless significant femoral deficit is present) is usually 75mm. However, the same thought processes need to be followed as per the tibial stem length. i.e. enough length to bypass poor distal bone but not too long to accelerate stress shielding.
In this case, the stem chosen was a universal slotted stem. Cemented stems are also available.
I would refer the viewer to the Op Tech regarding which laser etching corresponds to which stem length. The laster etching changes depending on stem thickness and type of femoral component used (PFC or TC3).

Sequentially ream until firm endosteal engagement is obtained.

Metaphyseal sleeves are also available for the femur. They were not used in this case however I show them for completeness.
The distal femur is expanded using the broach reamer and the trial stem.
If no sleeve is used and the chosen stem is less than 16mm in diameter, a stem reamer is used to expand the distal-most medullar which accommodates adapter.

If a sleeve is to be used, the sleeve broach and trial sleeve are attached and impacted. Again the broach is asymmetrical and ‘ANT’ should face anteriorly. The broach is impacted until the appropriate mark (A) on the handle (PFC or TC3) is level with the proposed distal femoral resection.
In my practice, I use the femoral sleeve infrequently. It is my opinion that there is additional stability gained from having a sleeve and it protects against rotational stresses and subsequent loosening. However I feel this is more important in the tibia and the box cut of the femoral implant also protects against the rotational torque forces.

Prepare the distal femoral cut.The cutting jigs will sit on the hand reamer. Use the last sized reamer where good endosteal bone resistance was encountered.
A supporting collar (A) is also used distally, the diameter of this matches that of the reamer.

The reamer is inserted until the appropriate laser-etched resection line is visible. The collar is impacted most distally (A). This adds additional stability and stops malalignment of the cutting jigs.

Confirm the set up of the distal cutting device for the distal femoral cutThe next four slides relate to making the distal femoral cut.
I have already covered this in-depth.
PFC Total knee replacement (De Puy-Synthes)
However, I will summarise again.
The distal locating device (A) is the starting block and the jig is constructed on this. The distal locating device sets the valgus angle (0-9 degrees), it is also sided (L/R). The degrees are set and the angle locked. My standard cut is 5 degrees, slightly lower in fixed valgus deformity knees and greater in fixed valgus knees or patients with short stature with ‘wide hips’ (i.e increased Q angle).
The outrigger (B) is placed in the distal locating device. The height is adjusted using button C. The distal cutting block (D) is placed on the outrigger. This is adjusted using the button (E).

Slide the distal locating device over the IM rod for the distal femoral cutEnsure there is no soft tissue interposition.

The cutting block should be set to zero or 1mm. The plan is for minor freshening cuts only.
Check with the angel wing that excess bone is not to be resected.
In cases of distal bone loss, the cutting block may be sitting on bone on the medial or lateral side only with a contralateral deficiency. The deficient side will require a distal augment. To perform a cut on the deficient side use the -4 or -8mm slots (seen in slide 41). A 4 or 8mm distal augment is then used in the final construct. It is important that these augments sit on bone.
Always be mindful that bone loss may have occurred on both medially and laterally and that distal augments are needed on both sides. Do not raise the joint line.
The joint line should be reviewed on the pre-operative radiographs.
Additionally, on table, the joint line can be approximated.
It is approximately 3cm or 2.5cm distal to the medial or lateral epicondyles respectively.
It is also approximately 1 fingerbreadth below the tip of the patella (leg in extension)
It lies at the level of any meniscal scars (if present).

Complete the ‘freshening’ cuts.
Once resected, healthy viable bone stock should be present, sufficient to ‘take the implant.’

Remove the distal cutting assembly and slide the A/P Chamfer cutting block over the IM nail.Confirm rotation. This step is important. There was no indication that pre-first stage, there was any problem with malrotation. Therefore an ‘angle wing’ through the cutting block resting on the anterior femoral surface confirms rotation. Confirmation of rotation can be sought by placing the knee at 90 degrees and adding a spacer block between the tibia and the AP / Chamfer block.
The cutting block is set to a 5 degree angle.

Confirm AP positioning of the block and then fix with drill pins. Both rotation and AP fit are important. I would match the revision femoral component size to the primary. Even so, positioning this is important. Place the component too posteriorly and the flexion gap may be too tight (and the femur may be notched). Place the component too anteriorly and the flexion gap is too loose. The position of the block relative to the IM rod can be adjusted. (see next slide).

The initial starting position of the cutting block is +2. (used in approx 85% of cases). This moves the block POSTERIORLY compared to the IM rod. (this always confuses me but relates to the STEM component being raised +2mm compared the femoral component).
0 and -2 move the block (and ultimately the femoral component) anteriorly, increasing the flexion gap.
The block is moved on the IM rod by inserting a hex screwdriver into the hex head (A), Position is read from the etched lines (B)
The setting for this case was -2mm

Complete the AP femoral cutsAgain at most these should be freshening cuts. However in cases where there has been bone loss, augments may be required, as always, the implant has to sit on good quality bone. As can be seen in the previous slide, additional cuts can be made at -4 and -8mm. There are no anterior augments.
I this case a posterior 8mm (medial) and 4mm (lateral) were required.

Complete the freshening anterior femoral chamfer cuts.

The correct size notch resection guide is selected.
The bushing used (A) corresponds the the AP settings dialed into the AP Chamfer block (ie. +2, 0 or -2mm) The bushing is also sided L/R so ensure the correct one is chosen.
If distal augments were used on the chamfer block they need to be used on the notch jig.
2 distal cuts are possible, where a PFC implant is used the more distal cut can be used (B). When using a TC3 femoral component, the notch is deeper (C). The cut is therefore proximal and in a slot proximal to the transverse bar (D)

The notch jig should sit on the anterior femoral cut, this will confirm orientation. It is held with pins. (Confirmation of rotation can be sought using the spacer blocks with the knee at 90 degrees if required).

The correct sized notch resection guide is applied to the cut femur and the femoral notch cut made.Remove the notch bushings and IM rod and make the notch cut. I prefer to use a reciprocating saw (https://www.orthoracle.com/library/sleeved-total-knee-replacement-tibial-plateau-fracture-mbt-depuy/operation/#slide-32)
However, an oscillating saw can be used with care.

Assemble the Femoral Trial. This is usually done by following the Op tech using the trials supplied.
The box trial is inserted into the femoral component (A).
The femoral adapter (B) is placed on the box and is secured with a stem bolt trial inserted distally to proximally through the notch of the femoral component. (see slide 70). This is tightened with a screwdriver.
The adapter needs to be set to the correct side (L/R) and position (+2 / 0/ -2mm which corresponds to the settings made on the AP / Chamfer block – C). In this instance, it is -2mm
The stem trial (D) is inserted and turned clockwise to secure.

Trial augments are attached. Both posterior condyles required augmentation.

Insert the femoral trial assembly

Ensure the construct is seating down well before impacting.
In very osteoporotic bone it is possible to impact the implant into any position. Therefore have it all correct before final firm impaction.

Confirm the tibia has remained seated before trialingCheck seating. The keel punch (A) is left in situ after impaction

Check that there is no soft tissue impinging prior to trial.
The keel punch (A) holds the trial MBT inserts.

Insert the MBT trial insert meniscus Once inserted check the usual movement range and balance.
The knee needs to be balanced both in extension and flexion, with no varus or valgus deformity or medial or lateral excessive tension. The patella should track normally. There should be no rotational malalignment.

Resurface the patellaThe original patella button was removed at the first stage revision. The patella cut is freshened again

The patella drill is used for the lug holes.

Insert the trial patella button.
Tracking should be reassessed.
Once this is all complete it is time to construct the definitive implants.
I would now take a look at the tourniquet time. If it is around 90 minutes I would release. It should not be up for more than 2 hours to prevent soft tissue (esp nerve) damage.
The tourniquet should be down 10 minutes for every hour it was inflated.
Complete haemostasis.

Complete assembly of the definitive Tibial Component Construction is all ex-vivo. The sleeve is constructed before the attachment of the stem.

As described earlier, there are two ways to attach the sleeve.
If I have been confident that the orientation of the sleeve and the tibial tray completely match then they can be constructed on-table as shown.
However, if I have been chasing a tibial defect and the sleeve and tibial tray are not completely aligned I will insert the sleeve into the bone, very gently insert the tibial tray in its correct alignment, tap the construct and remove from bone before final ex-vivo impaction (see slide-40 onwards, Sleeved Total Knee Replacement for Tibial Plateau Fracture (MBT DePuy)

Impact the sleeve on the tibial tray using the Sleeve Impactor
Warn the team that you are impacting. It makes a fair bit of noise, not good for the patient if surgery is not performed under GA.

The tibial stem is screwed home until hand-tight.

The stem is finally tightened using two wrenches. The first(A) grasps the sleeve, the second (B) tightens over the stem.

The final tibial construct ready for impaction

Prepare the definitive femoral construct.Where augments are going to be used the plastic plugs need to be removed. There is a sharp tool to do this on the kit. It is easier to do this before the implant is constructed.

Insert the bolt through the notch into the femoral adapter.
This bolt is either neutral (0) or +/-2.
For -2mm, the arrow should point down (posteriorly). This has taken the femoral adapter -2mm away from the anterior aspect of the femoral implant. This is in turn anteriorises the implant. (see the op tech diagrams – page 51, which make it clear).

Prior to final tightening check that the side settings (L/R) and position (+2 / 0/ -2) are correct. The markings are the same as the trials (the markings are best seen on the next slide).
The femur is held with the base adapter (A) which is tightened in place with the screw (B).
The torque wrench is placed over the femoral adapter and tightened until the deflection beam (C) reaches the laser etching.

When using 4mm and 8mm augments for both posterior and distal augmentation the distal augments should be inserted first.
Augments are screwed to the femoral component using ‘wobble bit’ adapters (A) and a torque screwdriver. I don’t know who names them.
The distal augments (not shown here) can go up to 16mm. The maximum for the PFC is 12mm for the TC2 – 16mm.
Posterior augments are smaller, for femoral sizes 1-3, 4mm is the maximum size. Size 4 and above 8mm posterior augments can be used. These obviously are guides only

The augments are screwed home using the wobble bits and a torque screwdriver. The bits are designed to snap off if too high torque is encountered.

Screwed home the femoral stem to hand tight.

The final tightening is accomplished using the base adapter and wrench. This is not on torque.
The kit is now ready for cementing.

Lavage, dry and then cement the tibiaThe in-depth cementing technique has been covered in my earlier presentation
Sleeved Total Knee Replacement for Tibial Plateau Fracture (MBT DePuy)
However, the cement is placed under pressure onto the plateau only. It does not go down into the metaphysis, this would inhibit bony on-growth onto the sleeve.

Implant first the tibial, then femoral and finally patella components.The tibial stem is inserted. Ensure correct alignment prior to final impaction.

Femoral impaction is very straightforward. Cement, again using the gun, is placed on the distal femur (none into the metaphysis) and the femoral construct impacted home.

Place the trial spacer in situ before cementing the patella.
The trial spacer needs to be inserted over the revision trial post. (see https://www.orthoracle.com/library/sleeved-total-knee-replacement-tibial-plateau-fracture-mbt-depuy/operation/#slide-47)

The tibial spacer consists of two parts. Firstly a trial revision post is inserted into the tray.

The trial spacer slotted onto this.
The knee held in extension (and very still) until the cement is cured.

Cement and compress the patella, removing any excess cement that extrudes.

Infiltrate deep tissues with local anaestheticI use bupivacaine but there are multiple variations on the amount and L/A mixes.
Finally lavage+++ Use litres, infection is a disaster.

Before inserting the definitive spacer check that there are no soft tissues which will impinge on the components

Slot home the MBT definitive spacer, check stability and range of motion, and close in layers.Once slotted home repeat the full range of movement exercise again checking for ROM, stability and patella tracking.

Complete a sound arthrotomy closure, and then close the wound in layers.Ensure the correct alignment of the arthrotomy during closure.

Complete fat layer closures

I use subcuticular monocryl for skin closure.

After the final dressing, I use a wool and crepe compression for a maximum of 24 hours.
I use an ice-wrap device in the post-operative period so the crepe and bandaging should not be in multiple layers. Early removal of all bandaging allows the ice-wrap to work more effectively.

Instead of showing the post op films I am going to show a number of x-rays that have been taken of knees with stems / sleeves. Some are mine, some a tertiary referrals.
The top row all failed, the bottom have not (so far!). Knowing what you now do from this presentation and the papers therein can you predict why the top row did not succeed when the bottom row has.